Microarray-Based Analysis of IncA/C Plasmid-Associated Genes from Multidrug-Resistant Salmonella enterica

Transcription

1 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Oct. 2011, p Vol. 77, No /11/$12.00 doi: /aem Copyright 2011, American Society for Microbiology. All Rights Reserved. Microarray-Based Analysis of IncA/C Plasmid-Associated Genes from Multidrug-Resistant Salmonella enterica Rebecca L. Lindsey,* Jonathan G. Frye, Paula J. Fedorka-Cray, and Richard J. Meinersmann U.S. Department of Agriculture, Agricultural Research Service, Bacterial Epidemiology and Antimicrobial Resistance Research Unit, Richard B. Russell Agricultural Research Center, 950 College Station Road, Athens, Georgia Received 11 March 2011/Accepted 28 July 2011 In the family Enterobacteriaceae, plasmids have been classified according to 27 incompatibility (Inc) or replicon types that are based on the inability of different plasmids with the same replication mechanism to coexist in the same cell. Certain replicon types such as IncA/C are associated with multidrug resistance (MDR). We developed a microarray that contains 286 unique 70-mer oligonucleotide probes based on sequences from five IncA/C plasmids: pyr1 (Yersinia ruckeri), ppip1202 (Yersinia pestis), pp (Photobacterium damselae), psn254 (Salmonella enterica serovar Newport), and pp91278 (Photobacterium damselae). DNA from 59 Salmonella enterica isolates was hybridized to the microarray and analyzed for the presence or absence of genes. These isolates represented 17 serovars from 14 different animal hosts and from different geographical regions in the United States. Qualitative cluster analysis was performed using CLUSTER 3.0 to group microarray hybridization results. We found that IncA/C plasmids occurred in two lineages distinguished by a major insertiondeletion (indel) region that contains genes encoding mostly hypothetical proteins. The most variable genes were represented by transposon-associated genes as well as four antimicrobial resistance genes (apha, merp, mera, and aada). Sixteen mercury resistance genes were identified and highly conserved, suggesting that mercury ion-related exposure is a stronger pressure than anticipated. We used these data to construct a core IncA/C genome and an accessory genome. The results of our studies suggest that the transfer of antimicrobial resistance determinants by transfer of IncA/C plasmids is somewhat less common than exchange within the plasmids orchestrated by transposable elements, such as transposons, integrating and conjugative elements (ICEs), and insertion sequence common regions (ISCRs), and thus pose less opportunity for exchange of antimicrobial resistance. The control of bacterial infections is threatened by the apparent increase in the number of bacteria that carry antimicrobial resistance (AR) genes (1). Plasmids are transmissible extrachromosomal genetic material that can carry AR genes (4, 7, 17). Other mobile elements responsible for the transfer of genetic material like AR genes include transposable elements such as transposons, integrating and conjugative elements (ICEs), and insertion sequence common regions (ISCRs) (5, 21, 37, 38, 40). In the family Enterobacteriaceae, plasmids have been classified into incompatibility (Inc) groups that are based on the inability of plasmids with the same replication mechanism to be maintained in the same cell lineage. In Enterobacteriaceae, there are 27 described Inc or replicon types (8). IncA/C plasmids are associated with multidrug resistance (MDR) (16, 39). We have previously found that IncA/C plasmids are prevalent in multidrug-resistant Salmonella enterica (23). Population genetics studies that have included data on the distribution of plasmids are common, but there have been very few studies on the distribution and flow of genes within a set of plasmids. Cataloging the genetic composition of plasmid populations can be used to make inferences about their history, which may be useful in understanding the selective pressures that have led to the prevalence of the plasmids and an appreciation of how anthropogenic effects may continue to influence their evolution. Such a study may not be meaningful with small plasmids, but a study of large plasmids may reveal how the plasmid evolved, accumulating and discarding various genes. IncA/C plasmids tend to be relatively large, generally 130 to 180 kb, with a complement of 160 to 210 genes (6, 16). Available DNA sequences indicate that there is heterogeneity in the complement of genes (6, 16). Welch et al. (39) presented data on the presence or absence of 12 loci based on PCR analysis of IncA/C plasmids from 51 isolates of S. enterica and 19 isolates of other members of the family Enterobacteriaceae, but they did not analyze the data for linkage disequilibrium. Here we used a previously described microarray with oligonucleotide probes for all the unique genes in five IncA/C plasmids that have been previously sequenced (24) to evaluate the gene content (286 loci) of IncA/C plasmids from a population of 59 different S. enterica isolates and to report on the core and accessory genome. We then used linkage disequilibrium analysis to examine the population genetics of these genes in these isolates. * Corresponding author. Mailing address: USDA-ARS-BEAR, Richard J. Russell Research Center, P.O. Box 5677, Athens, GA Phone: (706) Fax: (706) rebecca.lindsey@ars.usda.gov. Supplemental material for this article may be found at Published ahead of print on 12 August MATERIALS AND METHODS Bacterial isolates and plasmids. The 59 Salmonella enterica isolates in this study were selected from a diverse subset of 98 IncA/C-positive diagnostic isolates with known pulsed-field gel electrophoresis (PFGE) patterns (23). Diverse isolates were selected based on cluster analysis, antibiotic resistance, plasmid profile, source animal, and region. Cluster analysis revealed 48 unique PFGE types among the 59 isolates studied here, suggesting a variable population. These 6991

2 6992 LINDSEY ET AL. APPL. ENVIRON. MICROBIOL. isolates represented 17 serovars from 14 different sources, including canines (dogs) (n 4), cattle (n 17), chickens (n 3), dairy cattle (n 7), equines (horses) (n 6), felines (domestic cat) (n 1), reptiles (turtle and lizard) (n 3), swine (n 12), turkeys (n 3), mammals (alpaca and another mammal) (n 2), and an unknown source (n 1). Strains were obtained from all five regions in the United States as defined for the NARMS program ( 6750). Salmonella isolates were grown on brilliant green sulfa agar (BGSA) or blood agar plates (BAPs) at 37 C after recovering from frozen stock cultures stored at 70 C by standard methods ( 6750&page 1). DNA extraction and labeling. Total DNA from S. enterica isolates was extracted from 5 ml of overnight cultures grown in LB broth (Hardy Diagnostics, Santa Maria, CA) using the GenElute bacterial genomic DNA kit (Sigma, St. Louis, MO) following the manufacturers instructions for Gram-negative bacteria. DNA was labeled using random priming with N 15 (Operon, Huntsville, AL), and Cy3- or Cy5-dCTP (Amersham, Piscataway, NJ) was incorporated during extension with Klenow fragment (New England BioLabs, Beverly, MA) as previously described (28). DNA targets were purified following the manufacturer s directions using the QIAquick PCR purification kit (Qiagen, Valencia, CA) and eluted in 1 mm Tris-HCl (ph 8.0); the sample was then dried down to a volume of 20 l. Positive-control DNA from Yersinia ruckeri YR71 (39) was labeled with Cy3, while Salmonella enterica serovar Typhi CT18 (26), and negative-control DNA from Escherichia coli DH10B (11) was labeled with Cy5 and hybridized to arrays. The 59 study isolates were labeled with Cy3 or Cy5 and hybridized to arrays. Microarray construction, hybridization, and data analysis. The plasmid/antimicrobial resistance (PAR) microarray was constructed as previously described with 70-mer oligonucleotide probes for 286 unique IncA/C genes, 207 IncH1 plasmid genes, and 775 antimicrobial resistance genes (18, 24, 42) printed in triplicate. This study focused exclusively on the 286 unique IncA/C probes that were designed based on sequence obtained from the National Center for Biotechnology Information (NCBI) database ( from five IncA/C plasmids: pyr1 (Yersinia ruckeri strain YR71) (NCBI accession no. CP000602), pip1202 (Yersinia pestis biovar Orientalis strain IP275) (NCBI accession no. CP000603), pp (Photobacterium damselae subsp. piscicida) (NCBI accession no. AB277723), psn254 (Salmonella enterica subsp. enterica serovar Newport strain SL254) (NCBI accession no. CP000604), and pp91278 (Photobacterium damselae subsp. piscicida) (NCBI accession no. AB277724). The array was comprised of probes for all of the genes identified in pyr1 plus all the unique genes from the remaining plasmids. Protocols for hybridizations in formamide buffer were used for hybridization and the posthybridization wash processes as previously described (18, 19, 24, 28). The hybridized slides were scanned with a GenePix 4100A laser scanner (Axon, Foster City, CA) using a laser with a light wavelength of 532 nm or 635 nm to excite the Cy3 or Cy5 dye, respectively. Fluorescent images were captured as separate single-image tagged image file (TIF) format and analyzed with Scan- Array Express software version 1.1 (Packard BioChip Technologies). Hybridization signal intensities were measured, the means of the triplicate spots were recorded, and a cutoff two times the median of the mean hybridization intensity to all 70-mer probes was used. Microsoft Excel (Microsoft Corporation, Redmond, WA) was used to process hybridization results and to make comparisons of microarray hybridizations as previously described (19, 24). A positive hybridization to an oligonucleotide probe on the array was interpreted to indicate that the gene was present. Relationships between isolates were determined by hierarchical cluster analysis using open source CLUSTER 3.0 with Euclidean distance for gene content (10, 12). Statistical analysis. Linkage disequilibrium (LD) of probes on the array was calculated as an extension of Fisher s exact probability test on contingency tables (31) as instituted by the program Arlequin (13). Standard settings, such as 10,000 steps in the Markov chain and 1,000 dememorization steps, were used, and calculations of D (LD coefficient), D (maximum value standardized LD coefficient), and r 2 coefficients were made with a significance level of LD was not calculated for probes associated with hypothetical genes or genes that displayed all positive or all negative hybridizations. PCR detection of bla SHV-1. The bla SHV-1 gene was amplified using primers, conditions, and controls by the method of Rasheed et al. (29). RESULTS DNA isolated from 59 S. enterica isolates was hybridized to the plasmid/antimicrobial resistance (PAR) microarrays (23, 24). These isolates represented 17 serovars from 14 different host sources and were obtained from five regions of the United States (24). Fifty-nine different hybridization profiles were found. Hybridization results are illustrated in Fig. 1. Isolates are ordered across the top of Fig. 1 based on hierarchical cluster analysis that was performed with results of the microarray hybridizations. Analysis of the 180 oligonucleotide probes on the microarray that correspond specifically to the region 1 section of the array (the pyr1 genes; Fig. 1, region column, black or gray bar) reveals the presence of a region of insertions or deletions (indel) affecting 53 of those probes (pyr1x0087 to pyr1x0142). The first 18 isolates in Fig. 1 showed no or reduced hybridization (except for isolate 552) to the probes in the indel that consist mostly of probes that correspond to hypothetical protein genes (35 of 53 [66%]) and tracnuw, a single-stranded DNA-binding protein (ssb), assorted domain genes (cyclic diguanylate phosphodiesterase, von Willebrand factor type A, [2Fe-2S]-binding protein), and the insb1 transposase gene, to name a few. The remaining 41 isolates showed positive hybridizations to the probes in the indel. A second notable indel in region 1 of the array consists of 13 contiguous probes (pyr1x0147 to pyr1x0160) that correspond to genes for three hypothetical proteins, resolvase tnpr for transposon Tn5393, transposase insc for insertion sequence (IS) IS903, transposase insd for IS1133, transposase inse1 for IS10R, tetracycline resistance teta (class B), tetcd, transcriptional regulator (ArsR family), tetracycline repressor tetr (class B), and a sodium/glutamate symporter. Analysis of the remaining 127 oligonucleotide probes on the microarray in region 1 showed large regions of hybridizations (Fig. 1, region column, black bar); 114 of these probes bound to DNA from 53 of 59 (90%) or more of the isolates and 106 probes bound to DNA from 95% or more of the isolates. Region 2 (Fig. 1, region column, blue bar) of the microarray contains 40 oligonucleotide probes that correspond specifically to pip1202 genes. Analysis of region 2 (pip1202x0044 to pip1202x0198) showed most of these genes were present in two contiguous blocks: mercury resistance (8 oligonucleotide probes) and tetracycline resistance-associated regions. We also FIG. 1. Microarray hybridization results for this study. The leftmost column, the Name a column, lists the probe names, and the yellowhighlighted probes (listed top to bottom) correspond to the locations of the Welch et al. (39) primers along the IncA/C backbone. Column 2, the region b column, is color coded to indicate regions 1 to 5 on the array; region 1 with pyr1 probes is black or gray, region 2 with pip1202 probes is blue, region 3 with pp probes is red, region 4 with psn254 probes is yellow, and region 5 with pp91278 probes is orange. Columns 3 to 62 show hybridization results; blue indicates that the trait is present, and red indicates that the trait is absent. Isolate numbers are arranged across the top according to CLUSTAL 3.0 analysis. The last column, the Class c column, is the probe class, and the gene elements (6, 16) are color coded (mera [light blue], aada [light brown], beta-lactamase [light yellow], and flor [light pink] from Call et al. [6]).

3 VOL. 77, 2011 MICROARRAY ANALYSIS OF IncA/C PLASMID GENES 6993

4 APPL. ENVIRON. MICROBIOL. LINDSEY ET AL Fig. 1 Continued.

5 VOL. 77, 2011 MICROARRAY ANALYSIS OF IncA/C PLASMID GENES 6995 TABLE 1. IncA/C core genome based on positive hybridization of microarray probes Set or subset of isolates Total no. of genes detected Probe description (no. of positive probes representing genes or genetic elements with the same description) All 59 isolates (core genome) 42 Hypothetical protein (17), conserved hypothetical protein (7), putative membrane protein (4), putative lipoprotein (2), mercuric resistance mere (1), outer membrane lipoprotein blc1 (1), phosphoadenosine phosphosulfate reductase domain protein (1), putative parb partition (1), quaternary ammonium compound resistance suge1 (1), signal peptide peptidase sppa domain (1), staphylococcal nuclease domain protein (1), streptomycin resistance stra (1), transposase insb2 for insertion sequence (IS) IS26 (1), transposase insc3 for ISEc9 (1) 58 of 59 isolates 39 Hypothetical protein (14), conserved hypothetical protein (7), type IV conjugative transfer system traabgiv (5), transposase tnpa for transposon Tn21 and Tn5393 (2), beta-lactamase bla CMY-2 (1), dihydropteroate synthase sul2 (1), mercuric transport mert (1), Ner-like DNA-binding protein (1), putative membrane protein (1), putative para partition (1), putative type IV conjugative transfer system coupling factor (1), streptomycin resistance strb (1), tetracycline repressor, tetar class A (2), transposase insd1 for IS4321R (1) 57 of 59 isolates 25 Hypothetical protein (9), conserved hypothetical protein (5), type IV conjugative transfer system tradehl (4), DNA topoisomerase III (1), phage integrase (1), putative exonuclease (1), putative membrane protein (2), resolvase (1), transposase insa (1) 31 to 56 isolates of 59 isolates 89 Hypothetical protein (32), conserved hypothetical protein (15), 2Fe-2S -binding domain protein (1), alcohol dehydrogenase class III (1), alkylmercury lyase merb (1), aminoglycoside 3-Nacetyltransferase aac (1), aminoglycoside 3 -phosphotransferase (1), aminoglycoside adenylyltransferase aad (1), ATPase of the AAA family (1), C-5 cytosine-specific DNA methylase (1), cyclic diguanylate phosphodiesterase (EAL) domain (1), dihydropteroate synthase 1 sul (1), DNA replication terminus binding site (1), EAL domain urf2 (2), florfenicol/chloramphenicol resistance flor, Hg(II)-responsive transcriptional regulator merr (2), integrase inti1 for transposon Tn21 (1), mercuric reductase mera (1), mercuric resistance mere (3), mercuric resistance transcriptional repressor merd (2), mercuric transport mert (1), mercuric transport periplasmic component merp (2), modulator tnpm for transposon Tn21 (1), phage recombination bet (1), putative 5 -nucleotidase (1), quaternary ammonium compound resistance protein (1), protein family HMM PF00893 (1), Rhs family protein (1), single-stranded DNA-binding protein (1), transcriptional regulator (1), LysR family protein (1), transglycosylase SLT domain protein (1), transposase insc for IS903 (1), transposase inse, transposase tnpa for transposon Tn21 (1) found positive hybridizations to apha (aminoglycoside 3 phosphotransferase). Region 3 (Fig. 1, region column, red bar) of the microarray contains 13 oligonucleotide probes that correspond specifically to pp genes. Analysis of region 3 showed variability in an alcohol dehydrogenase and one of nine hypothetical proteins that were present. The oligonucleotide probe that corresponds to a transposase (P99018ORFx148) was not present in any of the isolates. Region 4 (Fig. 1, region column, yellow bar) of the microarray contains 47 oligonucleotide probes that correspond specifically to psn254 genes. Analysis of region 4 showed one contiguous block from psn254x0143 to psn254x0157, the aada region. Region 4 contains oligonucleotides for genes associated with aada and showed evidence of multiple, apparently unlinked insertions/deletions. Oligonucleotide probes associated with tetracycline repressor tetar, bla CMY-2, and tnpa for transposase Tn21 showed positive hybridizations in 98% of the isolates. Oligonucleotide probes associated with flor showed positive hybridizations in 90% of the isolates. Oligonucleotide probes associated with mercury resistance (mera), aminoglycoside (aada), transposases insef (ISCR2 and ISCR16) and insag, and transposition tnib showed the most variability in our population. Region 5 (Fig. 1, region column, orange bar) of the microarray contains 5 oligonucleotide probes that correspond specifically to pp91278 genes. Analysis of region 5 showed that three of four hypothetical proteins were present in 95% of the isolates. The oligonucleotide probe that corresponds to dihydrofolate reductase (NCBI accession no. YP_908609) was present only in isolate 858. A summary of positive probes is listed in Table 1. The first row lists the 42 positive probes for the IncA/C core genome of our isolates (i.e., found in all analyzed isolates). These include probes for genes encoding seven membrane proteins, including blc1, putative membrane proteins, a putative lipoprotein, and an outer membrane lipoprotein, two antimicrobial resistance genes (suge1 and stra), heavy metal resistance (mere), a signal peptide peptidase (sppa), two domain-containing proteins, partition (parb), and a few transposases. Twenty-four probes present in all isolates were of hypothetical function. The second row lists positive probes for 39 accessory genes (i.e., genes present in only 58 of 59 analyzed plasmids), including conjugation genes (traabgiv), antimicrobial resistance genes (bla CMY-2, tetar, and sulii were not present in isolate 527) (strb), 3 transposon-related genes, and 21 genes with hypothetical function (Table 1). Some of the 114 additional accessory genes (i.e., genes not present in 31 to 57 analyzed isolates) include conjugation genes (tradehl), antimicrobial resistance genes (aac, aad, flor, qac, strb, and suli), heavy metal resistance (merabdert), six transposon-related genes, and 61 genes with hypothetical function (Table 1). IncA/C plasmid core regions were described by Welch et al. and numbered 1 to 12 in numerical order around the plus strand of the IncA/C plasmid (39). We used our microarray hybridization results to detect the presence or absence of these core regions (Table 2). The majority of our plasmids (42 of 59 hybridizations) were positive for all of the probes.

6 6996 LINDSEY ET AL. APPL. ENVIRON. MICROBIOL. TABLE 2. Summary of positive hybridizations to IncA/C plasmid core regions as detected by microarray Isolate IncA/C core regions detected a n b c d e f Total on array 59 a IncA/C plasmid core regions are based upon the assay by Welch et al. (39), in which core regions are numbered in order 1 to 12 around the plus strand of the IncA/C plasmid sequence. Core regions are scored as present if the probes (highlighted in yellow in Fig. 1) were detected by microarray analysis of the isolate. The core region number is shown if the probe was detected by microarray analysis of the isolate., absent by microarray analysis. b n is the number of isolates with the same pattern. c Isolates 544, 517, 910, 897, 515, and 911 show the same pattern. d Isolates 604, 894, 518, 918, 915, and 573 show the same pattern. e Isolates 728 and 886 show the same pattern. f Isolates 552, 528, 674, 617, 587, 562, 845, 745, 514, 631, 930, 584, 519, 856, 633, 851, 752, 625, 550, 521, 621, 667, 600, 824, 934, 822, 563, 889, 812, 732, 733, 553, 548, 890, 888, 858, 589, 792, 784, 612, and 649 are positive for all 12 core regions. These regions are identified with the full hybridization data in Fig. 1. Pairwise linkage disequilibrium (LD) was calculated among all the probes on the array that were not hypothetical or invariantly positive or negative (see Fig. S1 in the supplemental material). The PAR3 array contained 1,395 (22.4%) probe pairs positive for linkage at P values of 0.05 out of the total of 6,216 pairs that were analyzed. The PAR3 array contained nine pairs of antimicrobial gene probes that were contiguous, eight (88.9%) of which showed linkage disequilibrium; 162 pairs were not contiguous, and 26 (16%) of these pairs were linked. The PAR3 array contained 35 pairs of conjugation gene probes that were contiguous, 23 (66%) of which were linked; 85 pairs were not contiguous, and 42 (49%) of these pairs were linked. The PAR3 array contained 25 pairs of transposon gene probes that were contiguous, 11 (44%) of which were linked; 132 pairs were not contiguous, and 23 (17%) of these pairs were linked. The PAR3 array contained 3 pairs of DNA-binding protein gene probes that were contiguous, all 3 (100%) of which were linked; 63 pairs were not contiguous, and 6 (9.5%) of these pairs were linked. DISCUSSION The objective of this study was to learn more about the population genetics of IncA/C plasmids by characterizing a diverse population of 59 S. enterica isolates from different animal sources. Isolates were hybridized to a plasmid/antimicrobial resistance (PAR) hybridization array with 286 unique oligonucleotide plasmid probes from IncA/C plasmids, and the results were analyzed to determine the core genome and accessory genome in our isolates (Fig. 1 and Table 1). The universally present core gene set appeared to be very limited in expected plasmid maintenance-related genes; parb was the only core gene recognized as needed for replication, but there were 30 genes of unknown function in the core set, some of which are likely to be used in plasmid maintenance. The accuracy of the PAR hybridization array was previously analyzed (24) with control strains S. Typhi CT18 (26), Y. ruckeri strain YR71 (39), and S. Typhimurium LT2 (25). All the probes on the array originated from IncA/C sequences; however, it is possible for chromosomal DNA from the tested strains to give a positive hybridization result to some of the IncA/C-associated probes. Cluster analysis of the oligonucleotide probes on the microarray showed two lineages of IncA/C plasmids that differed mainly in the presence of an insertion, deletion, or combination of the two (indel) of 53 (pyr1x0087 to pyr1x0142) of the 180 oligonucleotide probes that correspond to region 1 (Fig. 1, region column, black bar). The lineage that consists of isolates 928 to 915 (Fig. 1, top row) was 30% (18 of 59) of the total isolates and did not hybridize to approximately 53 oligonucleotide probes from the pyr1 backbone section of the array. Interestingly, this indel occurred at the same location as an integrating and conjugative element (ICE) hot spot 2 (HS2) (40). The indel in the lineage did not hybridize to the probe for the single-strand DNA-binding protein gene (ssb); in previous findings, the sequence for ssb was present in pyr1, psn254, pip1202, p99-018, and p91278 but missing in pra1 (16). Because the indel was not identical in all the plasmids, it is likely that it was shaped by more than one genetic event. Welch et al. used 12 loci along the pyr1 IncA/C backbone for a PCRbased screening assay (Fig. 1, loci highlighted in yellow) to investigate the core backbone of IncA/C plasmids (39). In this study, we found an indel dividing our IncA/C plasmids into two lineages in which loci 6 to 9 (pyr1x0087 to pyr1x0142) had no or reduced hybridization in one lineage of 17 isolates, and loci 6 to 9 were present in the second lineage of 42 isolates (Fig. 1 and Table 2) (39). In the lineage of 17 isolates with no or reduced hybridization, six isolates were negative for the core regions encompassed by Welch primers 6 to 9 (Fig. 1, loci highlighted in yellow), another six isolates were negative for core regions 7 to 9, and the five remaining isolates were negative for at least one locus in core regions 7 to 9 (Table 2). Core regions 6 through 9 are within a previously defined insertion/deletion (indel) for IncA/C plasmids and are in the same location as an ICE HS2. Forty-one isolates had genes detected in all 12 core regions located around the genome of IncA/C plasmids (Fig. 1 and Table 2). Welch et al. examined 51 IncA/C-positive Salmonella isolates recovered from retail meats from 2000 to 2005, and 16% of these isolates did not

7 VOL. 77, 2011 MICROARRAY ANALYSIS OF IncA/C PLASMID GENES 6997 contain loci 6 through 9. Here the lineage without Welch loci 6 to 9 by microarray hybridizations comprises 30% of our population. The differences in the two studies could be due to the different isolate sources (retail meats versus veterinary clinical isolates from sick animals) or time frame (2000 to 2005 versus 2005). In region 1 (Fig. 1), there were 16 contiguous probes (pyr1x0148 to pyr1x0163) associated with a combined Tn10 and Tn5393 transposon cassette previously found in IncA/C plasmids (15, 39). In this study, three genes associated with the Tn5393 transposon were highly conserved, 98% (58 of 59) of our isolates showed positive hybridization to tnpa for Tn5393 (pyr1x0148) and to stra and strb (pyr1x0162 and pyr1x0163). These genes are part of the Tn5393 transposon cassette that confers aminoglycoside resistance and that has widespread dissemination (34). The Tn5393 cassette has been found in the United States associated with the apple and pear tree pathogen Erwinia amylovora (9), in Italy from multidrugresistant S. enterica of mostly poultry origin (27), in Norway in the fish pathogen Aeromonas salmonicida (20), and in Pseudomonas syringae and Xanthomonas campestris (35). The remaining 127 oligonucleotide probes in region 1 of the microarray showed large regions of similarity among the 59 isolates hybridized. In addition to plasmids, there are other mechanisms for the transfer of genes responsible for antibiotic or heavy metal resistance, including transposable elements, such as transposons and their corresponding integrons, ICEs, and insertion sequence common regions (ISCRs) (5, 21, 36 38, 40). Probes on the array associated with transposable elements like transposons, ICEs, and ISCRs showed the most variability. Analysis showed that the four antimicrobial or heavy metal resistance probes on the array with the greatest variability were aminoglycoside 3 -phosphotransferase (apha, region 2, pip1202x0052), mercuric transport protein periplasmic component (merp, region 2, pip1202x0159), mercuric reductase (mera, region 2, pip1202x0161), and aminoglycoside adenyltransferase (aada, region 4, psn254x0144). All four probes had significant linkage to transposon-related probes (see Fig. S1 in the supplemental material). Aminoglycoside 3 -phosphotransferase (apha) provides resistance to kanamycin-type aminoglycoside antibiotics (30). Hybridizations to this array showed 54% (32 of 59) of the isolates with positive hybridization to the apha (pip1202x0156) probe from Y. pestis. Aminoglycoside adenyltransferase (aada, psn254x0144), which confers resistance to streptomycin, is within an integron designated In2 (32) and is part of the Tn21 transposon family (21). Linkage disequilibrium (LD) analysis of our microarray results found that aada was significantly linked to the five Tn21-related probes on the array. In psn254, the aada region includes the aada, aacc, qac, sul1, gros, and grol genes (6, 16, 39) and had a high degree of variability in our isolates. This array contains two distinct variants of mer operons that contain merp, in region 2 from pip1202 (Y. pestis biovar Orientalis strain IP275) and another from region 4 from psn254 (S. enterica subsp. enterica serovar Newport strain SL254). Both mer operons contain the basic set of Gram-negative mer genes, merrtpade, and are associated with transposons. The Y. pestis probe set also contains the mercuric transporter merc, which is redundant to mert. Here we found merc less frequently than mert, which corresponds with the results from an earlier study (22). The S. enterica probe set contains the organomercurial lyase MerB and the corresponding organomercurial responsive version of MerR. It is interesting to note that the mercury resistance genes on the IncA/C backbone were highly conserved in both lineages. The periplasmic MerP protein (merp, pip1202x0159) and the cytosolic mercuric reductase (mera, pip1202x0161) have frequently been found in IncA/C backbones (6). We found significant linkage disequilibrium between mera or merp probes and tetracycline resistance- and transposon-related probes (see Fig. S1 in the supplemental material). In a phylogenetic study of 213 bacterial mera sequences, Barkay et al. found that bacterial and plasmid MerA loci were often associated with insertion elements and integrase or transposase genes (3). Barkay et al. examined MerA evolution and concluded that MerA originated in a thermophilic bacterial lineage after the divergence of the Archaea and Bacteria domains from the most-recent common ancestor approximately 2.4 billion years ago (3). Mercury resistance could have evolved in the microflora of animals living in or near natural mercury deposits. However, resistance to other heavy metals may be the driving force for selection of these genes. The selective pressure for mercury resistance in humans and animals like the ones in the present study could be attributed to the use of mercury as a therapy for parasites and syphilis from the 17th century to the early 20th century and the current exposure of humans to mercuric mercury from silver amalgam dental fillings which is continuously leached into the saliva and gut contents of humans with these fillings (33). Currently, 35 metric tons of mercury is used each year in the United States in dental fillings. It is interesting to note the positive hybridization of probes inse (ISCR element 2 [ISCR2]) and insf (ISCR16) in this study that were present in 34 and 4 isolates, respectively (Fig. 1, region 4). ISCRs include elements capable of antibiotic resistance gene capture and movement and can construct clusters of antibiotic resistance genes on chromosomes and in plasmids (36 38). Toleman et al. noted that ISCR16 had been found in only three IncA/C plasmids at that time (38); here we found it present in four additional IncA/C plasmids. LD analysis of ISCR2 and ISCR16 revealed significant linkage to transposonrelated mercury resistance and antibacterial resistance probes (see Fig. S1 in the supplemental material). All of the S. enterica isolates in this study except for isolate 527 showed a positive hybridization to the beta-lactamase bla CMY-2 probe. More than one copy of bla CMY is frequently found in S. enterica isolates (6, 39); however, this microarray does not distinguish how many copies of each gene is present, it can only distinguish the presence or absence of hybridization to an oligonucleotide that is representative of a gene. Isolate 888 displayed a positive hybridization to the beta-lactamase bla SHV-1 located in region 2 of the array (NCBI accession number YP_ ). This isolate was assayed by PCR, because bla CMY and bla TEM have been associated with betalactam-resistant isolates from S. enterica (16, 39, 41), while bla SHV is found in Klebsiella pneumoniae (2, 14). The assay failed to detect bla SHV, indicating a false-positive hybridization possibly due to the bla SHV probe cross hybridizing with other detected genes. Hybridization analysis of the 47 oligonucleotide probes in region 4 of the array displayed the most variability of all re-

8 6998 LINDSEY ET AL. APPL. ENVIRON. MICROBIOL. gions of the array (Fig. 1). This is interesting because DNA hybridized to the array was extracted from S. enterica isolates, which is presumably more closely related than the sequences from the other species. Region 4 of the microarray contains oligonucleotides of genes associated with mera, aada, flor, and transposases. As expected, oligonucleotide probes associated with tetracycline repressor proteins TetA and TetR and with bla CMY-2 and TnpA for transposase Tn21 were highly conserved with positive hybridizations in 98% of the isolates in this study. LD analysis revealed that linkage for all groups (antibacterial, conjugation, transposon, and DNA binding) was enriched if the genes were contiguous. Conjugation genes that were not contiguous also showed enrichment for linkage (Fig. 1; see Fig. S1 in the supplemental material). This may reflect selective pressures from the functional properties of the gene products. In summary, we found that the IncA/C core genome of our isolates (i.e., positive hybridizations in all analyzed isolates) consists of 42 genes of which 24 were for genes encoding hypothetical proteins. IncA/C plasmids occur in two lineages separated by a major insertion-deletion (indel) that contains mostly genes encoding hypothetical proteins. The most variable genes were represented by transposable element-like transposons, ISCRs, and ICEs as well as four antimicrobial or mercury resistance genes. Mercury resistance genes are highly conserved in these IncA/C plasmids. Our previous study showed that the associations of Salmonella species, IncA/C plasmids, and multidrug resistance (MDR) genes are very old (23). Mercury resistance is also very old (3) and has been associated with plasmids and transposable elements for a long time, so it is not surprising that they have been found together on IncA/C plasmids. On the basis of differences in linkage disequilibrium, our studies suggest that the transfer of antimicrobial resistance determinants by transfer of IncA/C plasmids is somewhat less common than exchange within the plasmids orchestrated by transposable elements. Thus, the IncA/C plasmids have been remodeled by transposons more often than the plasmid was exchanged with other bacterial lineages. The donors of the transposons remain unknown, but they may be other plasmids or conjugative units that are more transient in the bacterial population. ACKNOWLEDGMENTS We gratefully acknowledge Anne O. Summers for critically reading the manuscript and for her expertise and input on the mercury resistance operon. We thank Linda Genzlinger, LaShanda Glenn, Tiffanie Woodley, and Tyler Wilcher for technical assistance. The mention of trade names or commercial products in this manuscript is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. REFERENCES 1. Alanis, A. J Resistance to antibiotics: are we in the post-antibiotic era? Arch. Med. Res. 36: Babini, G. S., and D. M. Livermore Are SHV beta-lactamases universal in Klebsiella pneumoniae? Antimicrob. Agents Chemother. 44: Barkay, T., K. Kritee, E. Boyd, and G. Geesey A thermophilic bacterial origin and subsequent constraints by redox, light and salinity on the evolution of the microbial mercuric reductase. Environ. Microbiol. 12: Bennett, P. M Plasmid encoded antibiotic resistance: acquisition and transfer of antibiotic resistance genes in bacteria. Br. J. Pharmacol. 153(Suppl. 1):S347 S Burrus, V., G. Pavlovic, B. Decaris, and G. Guedon Conjugative transposons: the tip of the iceberg. Mol. Microbiol. 46: Call, D. R., et al bla CMY-2 -positive IncA/C plasmids from Escherichia coli and Salmonella enterica are a distinct component of a larger lineage of plasmids. Antimicrob. Agents Chemother. 54: Carattoli, A Importance of integrons in the diffusion of resistance. Vet. Res. 32: Carattoli, A Resistance plasmid families in Enterobacteriaceae. Antimicrob. Agents Chemother. 53: Chiou, C. S., and A. L. Jones Nucleotide sequence analysis of a transposon (Tn5393) carrying streptomycin resistance genes in Erwinia amylovora and other Gram-negative bacteria. J. Bacteriol. 175: de Hoon, M. J., S. Imoto, J. Nolan, and S. Miyano Open source clustering software. Bioinformatics 20: Durfee, T., et al The complete genome sequence of Escherichia coli DH10B: insights into the biology of a laboratory workhorse. J. Bacteriol. 190: Eisen, M. B., P. T. Spellman, P. O. Brown, and D. Botstein Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. U. S. A. 95: Excoffier, L., G. Laval, and S. Schneider Arlequin version 3.0: an integrated software package for population genetics data analysis. Evol. Bioinform. Online 1: Ford, P. J., and M. B. Avison Evolutionary mapping of the SHV beta-lactamase and evidence for two separate IS26-dependent blashv mobilization events from the Klebsiella pneumoniae chromosome. J. Antimicrob. Chemother. 54: Fricke, W. F., et al Antimicrobial resistance-conferring plasmids with similarity to virulence plasmids from avian pathogenic Escherichia coli strains in Salmonella enterica serovar Kentucky isolates from poultry. Appl. Environ. Microbiol. 75: Fricke, W. F., et al Comparative genomics of the IncA/C multidrug resistance plasmid family. J. Bacteriol. 191: Frost, L. S., R. Leplae, A. O. Summers, and A. Toussaint Mobile genetic elements: the agents of open source evolution. Nat. Rev. Microbiol. 3: Frye, J. G., et al DNA microarray detection of antimicrobial resistance genes in diverse bacteria. Int. J. Antimicrob. Agents 27: Frye, J. G., et al Development of a DNA microarray to detect antimicrobial resistance genes identified in the National Center for Biotechnology Information database. Microb. Drug Resist. 16: L Abee-Lund, T. M., and H. Sorum Functional Tn5393-like transposon in the R plasmid pras2 from the fish pathogen Aeromonas salmonicida subspecies salmonicida isolated in Norway. Appl. Environ. Microbiol. 66: Liebert, C. A., R. M. Hall, and A. O. Summers Transposon Tn21, flagship of the floating genome. Microbiol. Mol. Biol. Rev. 63: Liebert, C. A., A. L. Watson, and A. O. Summers The quality of merc, a module of the mer mosaic. J. Mol. Evol. 51: Lindsey, R. L., P. J. Fedorka-Cray, J. G. Frye, and R. J. Meinersmann IncA/C plasmids are prevalent in multidrug-resistant Salmonella enterica isolates. Appl. Environ. Microbiol. 75: Lindsey, R. L., J. G. Frye, P. J. Fedorka-Cray, T. J. Welch, and R. J. Meinersmann An oligonucleotide microarray to characterize multidrug resistant plasmids. J. Microbiol. Methods 81: McClelland, M., et al Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature 413: Parkhill, J., et al Complete genome sequence of a multiple drug resistant Salmonella enterica serovar Typhi CT18. Nature 413: Pezzella, C., A. Ricci, E. DiGiannatale, I. Luzzi, and A. Carattoli Tetracycline and streptomycin resistance genes, transposons, and plasmids in Salmonella enterica isolates from animals in Italy. Antimicrob. Agents Chemother. 48: Porwollik, S., J. Frye, L. D. Florea, F. Blackmer, and M. McClelland A non-redundant microarray of genes for two related bacteria. Nucleic Acids Res. 31: Rasheed, J. K., et al Evolution of extended-spectrum beta-lactam resistance (SHV-8) in a strain of Escherichia coli during multiple episodes of bacteremia. Antimicrob. Agents Chemother. 41: Shaw, K. J., P. N. Rather, R. S. Hare, and G. H. Miller Molecular genetics of aminoglycoside resistance genes and familial relationships of the aminoglycoside-modifying enzymes. Microbiol. Rev. 57: Slatkin, M Linkage disequilibrium in growing and stable populations. Genetics 137: Stokes, H. W., and R. M. Hall A novel family of potentially mobile DNA elements encoding site-specific gene-integration functions: integrons. Mol. Microbiol. 3: Summers, A. O., et al Mercury released from dental silver fillings provokes an increase in mercury- and antibiotic-resistant bacteria in oral and intestinal floras of primates. Antimicrob. Agents Chemother. 37: Sundin, G. W Distinct recent lineages of the stra-strb streptomycin-

9 VOL. 77, 2011 MICROARRAY ANALYSIS OF IncA/C PLASMID GENES 6999 resistance genes in clinical and environmental bacteria. Curr. Microbiol. 45: Sundin, G. W., and C. L. Bender Expression of the stra-strb streptomycin resistance genes in Pseudomonas syringae and Xanthomonas campestris and characterization of IS6100 in X. campestris. Appl. Environ. Microbiol. 61: Toleman, M. A., P. M. Bennett, and T. R. Walsh ISCR elements: novel gene-capturing systems of the 21st century? Microbiol. Mol. Biol. Rev. 70: Toleman, M. A., and T. R. Walsh Evolution of the ISCR3 group of ISCR elements. Antimicrob. Agents Chemother. 52: Toleman, M. A., and T. R. Walsh ISCR elements are key players in IncA/C plasmid evolution. Antimicrob. Agents Chemother. 54: Welch, T. J., et al Multiple antimicrobial resistance in plague: an emerging public health risk. PLoS One 2:e Wozniak, R. A., et al Comparative ICE genomics: insights into the evolution of the SXT/R391 family of ICEs. PLoS Genet. 5:e Zhao, S., et al Lactam resistance in Salmonella strains isolated from retail meats in the United States by the National Antimicrobial Resistance Monitoring System between 2002 and Appl. Environ. Microbiol. 75: Zou, W., et al Microarray analysis of antimicrobial resistance genes in Salmonella enterica from preharvest poultry environment. J. Appl. Microbiol. 107: